3C.6
AFRICAN EASTERLY WAVE CONVECTION AND TROPICAL
CYCLOGENESIS: A LAGRANGIAN PERSPECTIVE
Kenneth D. Leppert II * and Daniel J. Cecil
University of Alabama in Huntsville, Huntsville, Alabama
1. INTRODUCTION
to one day for this study]) for June – November of
2001–2010 over a region stretching from 130°W –20°E
(from the East Pacific to West Africa), 5°–20°N.
After the various wave phases were identified,
information from National Hurricane Center (NHC) storm
reports (NHC 2011) was utilized to identify AEW troughs
that were associated with the development of a tropical
cyclone of at least tropical storm strength (i.e.,
developing waves; DWs). NHC (2011) was also used to
determine the day on which a tropical depression was
first identified (i.e., day zero; D0) for each DW. Then
each wave trough was traced up to five days prior to D0
(D-5) until one day after D0 (D+1). Any of the other
three wave phases found within three data points (7.5°)
east or west of the DW troughs were considered to be
part of the DW.
In addition, we sought to compare the evolution of
DWs with that of waves that never developed a tropical
cyclone (i.e., nondeveloping waves; NDWs). Day zero
for NDWs was defined as the day on which these waves
achieved a maximum in 850-hPa vorticity within the
trough phase.
Presumably, the strongest, most
coherent NDWs are most relevant for cyclogenesis.
Hence, only those NDWs that could be tracked for at
least seven days (NDWs were tracked manually using
Hovmoller diagrams) and achieved a maximum 850-hPa
-1
vorticity of at least 8.0E-06 s (equal to the mean 850hPa vorticity of all NDW troughs plus one standard
deviation; for comparison, the mean 850-hPa trough
-1
vorticity of DWs on D0 is 5.1E-06 s ) were included in
the analysis. Once D0 for NDW troughs was identified,
other days relative to D0 as well as other wave phases
were identified for NDWs as they were for DWs.
Using a similar methodology to that used by
Hopsch et al. (2010), composites of several large-scale
variables, including θe and vorticity, were created as a
function of wave phase and day relative to D0 for all
DWs and NDWs using data taken from the NCEPNCAR reanalysis dataset. These composites were
used to assess the evolution of the large-scale wave
circulation and structure.
In order to assess the evolution of the coverage by
cold cloudiness associated with easterly waves, the 4km IR brightness temperatures from the NASA globalmerged IR brightness temperature dataset (Liu et al.
2009) were utilized. In particular, the IR brightness
temperatures were used to find the fraction of each 2.5°
box covered by temperatures ≤210 K and ≤240 K.
These areal coverage fractions were subsequently
composited similar to the large-scale reanalysis
variables.
Several different datasets were used to assess the
intensity of convection and its associated evolution
including lightning data from the lightning imaging
sensor (LIS; Boccippio et al. 2002; flash rates valid for
African easterly waves (AEWs) have been shown to
be important for tropical cyclogenesis in the Atlantic
(e.g., Landsea 1993) and East Pacific (e.g., Avila 1991;
Molinari and Vollaro 2000). In particular, these waves
often provide a precursor low-level disturbance
necessary for tropical cyclogenesis (Kurihara and
Tuleya 1981). In addition, convection resident in AEWs
may play an important role in tropical cyclogenesis via
thermodynamic and dynamic feedbacks to the larger
meso-to-synoptic scale circulation. For example, more
widespread and/or intense convection could act to
moisten mid and upper levels as a result of transport in
convective updrafts.
Strong, evaporatively-cooled
downdrafts that can transport lower values of equivalent
potential temperature (θe) towards the surface may be
inhibited by this mid- to upper-level moistening, thus
aiding tropical cyclogenesis (Rotunno and Emanuel
1987). By inhibiting strong downdrafts, the increase of
mid- and upper-level moisture could also help to
intensify updrafts and convection (e.g., Nolan 2007).
This enhanced upward motion could aid the
intensification and/or development of a mid- to low-level
circulation (Nolan 2007; Raymond et al. 2011) via the
stretching term in the vorticity equation. Thus, an
enhancement of vorticity is another way convection
could potentially aid tropical cyclogenesis.
The purpose of this study is to use a Lagrangian
framework to analyze the evolution of and relation
between AEWs, convection, and tropical cyclogenesis.
In particular, we seek to determine the evolution of
several convective characteristics (e.g., coverage and
intensity) and the corresponding evolution of the
thermodynamics and dynamics of larger-scale waves in
the days leading up to and including tropical
cyclogenesis. In the next section, a description of the
methodology used is provided. Section 3 provides the
main results from the study, and conclusions are given
in section 4.
2. METHODOLOGY
Following the methodology of Leppert and Petersen
(2010) easterly waves were analyzed and separated
into ridge, northerly, trough, and southerly phases using
NCEP-NCAR reanalysis (Kalnay et al. 1996) 700-hPa
meridional wind data (note that the reanalysis has a
spatial [temporal] resolution of 2.5° [six hours; averaged
____________________________________________
* Corresponding author address: Kenneth D. Leppert II,
NSSTC, 320 Sparkman Dr., Rm 4074, Huntsville, AL
35805; e-mail: leppert@nsstc.uah.edu.
1
Table 1. The fraction of a 2.5° by 2.5° grid box (76176
2
km ) covered by IR brightness temperatures less than or
equal to 240 K as a function of wave phase and day
relative to D0 for developing and nondeveloping waves
(a land mask has been applied to both wave types).
The values are valid from five days prior to D0 (D-5) to
one day after D0 (D+1). The bold numbers indicate
values that are statistically significantly (at the 99%
level) greater than the corresponding values of the other
wave category based on the analysis of variance
technique, and italics indicate values that are
significantly greater than the value from the previous
day from the same wave type and phase.
each day and 2.5° grid box were calculated and then
composited similar to other variables) on board the
Tropical Rainfall Measurement Mission (TRMM)
satellite, radar reflectivity data from the TRMM
precipitation radar (PR; Kummerow et al. 1998), and
37.0/85.5-GHz brightness temperatures from the TRMM
microwave imager (TMI; Kummerow et al. 1998).
3. RESULTS
Some waves may have developed a tropical
cyclone and/or achieved their maximum low-level
vorticity close to the West African coast or the west
coast of Central America. Hence, the evolution of some
easterly waves from D-5 to D0 includes a transition from
over land to ocean. Because convective intensity and
lightning generally decrease over the ocean compared
to land (e.g., Toracinta et al. 2002; Christian et al.
2003), it is possible that a decrease in the intensity of
convection may be observed for waves as they evolve
from D-5 to D0 unrelated to the evolution towards
cyclogenesis. Therefore, a land mask was applied to all
composites.
The composite coverage by IR brightness
temperatures ≤240 K provided in Table 1 (as a function
of wave phase and day relative to D0) indicate that the
trends in coverage by cold cloudiness vary as a function
of phase in the days leading up to tropical cyclogenesis
for DWs. For example, the coverage increases from D5 until D-2 in the ridge and northerly phases and then
decreases thereafter.
In contrast, the coverage
generally increases from D-5 to D0 in the trough and
southerly phases (note that coverage decreases on D+1
in these wave phases). The IR threshold values for
NDWs provided in Table 1 suggest an evolution of the
coverage by cold cloudiness that is similar to that of
DWs. In particular, the NDW ridge phase shows much
variability in coverage from D-5 to D+1 while the other
phases show a general increasing trend on all days
except D+1. While the evolution of coverage by cold
cloud tops is similar between DWs and NDWs, the
magnitude of values are significantly greater (i.e.,
statistical significance at the 99% level using the
analysis of variance technique) for DWs on nearly all
days and in all phases, except the ridge phase. Hence,
it appears that a greater coverage by cold cloudiness is
more relevant to tropical cyclogenesis than the evolution
of that coverage in the days leading up to genesis.
In contrast to what is observed for the coverage by
cold cloudiness (especially in the trough and southerly
phases), the intensity of DW convection as indicated by
composite lightning flash rates (Fig. 1a) appears to
decrease as D0 is approached. Specifically, both the
trough and southerly phases show a decrease in flash
rate every day after D-4. The northerly and ridge phase
values vary from day to day, but flash rates in both
phases decrease by more than 60% from D-5 until D0.
The LIS flash rates of NDWs shown in Fig. 1b show
much variability from day to day with little indication of
any consistent trends. However, the flash rates of all
NDW phases are greater on D0 and D+1 when
compared to the corresponding values on D-5,
Developing Waves
Day
Ridge
Northerly
Trough
Southerly
D-5
0.094
0.100
0.112
0.090
D-4
0.102
0.107
0.122
0.112
D-3
0.105
0.124
0.121
0.118
D-2
0.107
0.137
0.146
0.143
D-1
0.102
0.132
0.152
0.170
D0
0.096
0.114
0.171
0.177
D+1
0.088
0.100
0.163
0.161
Nondeveloping Waves
D-5
0.068
0.054
0.076
0.096
D-4
0.064
0.066
0.067
0.103
D-3
0.067
0.075
0.070
0.084
D-2
0.100
0.093
0.098
0.094
D-1
0.112
0.105
0.125
0.107
D0
0.103
0.117
0.139
0.123
D+1
0.096
0.082
0.113
0.108
suggesting an increase in intensity with time for these
waves or, at least, a constant intensity with time (it is
possible that more flashes are recorded by LIS due to
the increasing coverage by [potentially electricallyactive] cold cloudiness with time, giving the impression
of an increasing flash rate with time while the flash rate
[i.e., intensity] of individual convective elements does
not change), in contrast to DWs which show a decrease
in lightning with time despite an increase in coverage of
cold cloudiness. Note, however, that any trends in
lightning flash rates observed for either DWs or NDWs
are small compared to the standard deviation of 144.5
-1
flashes day .
Another difference between DWs and NDWs can
be observed in the evolution of large-scale profiles of
the θe anomaly shown in Fig. 2 as a function of day
relative to D0 for the trough phase. The DWs show a
consistent increase in moisture throughout a large depth
of the troposphere on nearly all days, while NDWs only
show an increase on some days over a relatively
2
Fig. 2. Composite analysis of vertical profiles of the θe
anomaly (anomaly relative to the mean at each pressure
level and longitude) as a function of day relative to D0
for the trough phase (a land mask has been applied).
The black (blue) lines are for developing waves
(nondeveloping waves). The profiles are valid for five
days prior to D0 (D-5) to one day after D0 (D+1; note
that the D-4 and D-2 profiles are not shown for clarity),
and the horizontal dashed lines indicate the value of the
standard deviation at each level.
1000 hPa on D0 and D+1 (although, these increases
are admittedly small). Note that the initial increase in
midlevel vorticity for DWs appears to be consistent with
several previous studies (e.g., Nolan 2007; Raymond et
al. 2011). In contrast, NDWs show an increase in
cyclonic vorticity below 500 hPa on D-1 and D0 with a
subsequent decrease on D+1. Note in particular the
large increase in 850-hPa vorticity for NDWs on D0,
which, of course, is due to how D0 is defined for NDWs.
At upper levels (~200 hPa), DWs exhibit an increase in
anticyclonic vorticity on all days except D+1 (NDWs are
associated with an increase in upper-level anticyclonic
vorticity until D-1 that remains nearly constant
thereafter). The development of greater upper-level
anticyclonic vorticity for DWs combined with the
increase in mid- and low-level cyclonic vorticity suggests
the development of a warm core in association with
thermal wind balance, consistent with tropical
cyclogenesis.
Fig. 1. Composite lightning flash rates as a function of
day relative to D0 for each wave phase of a.) developing
waves and b.) nondeveloping waves after application of
a land mask. Note that the labels along the abscissa
are for five days prior to D0 (D-5) to one day after D0
(D+1). The thin, horizontal lines indicate the average
flash rate over the full analysis domain, and the
-1
standard deviation is 144.5 flashes day .
shallow depth. In addition, the θe anomaly is greater for
DWs compared to NDWs on nearly all days and levels.
Some differences are also observed between DWs
and NDWs in terms of the evolution of the large-scale
wave circulation as shown in Fig. 3 by composite
profiles of relative vorticity for the trough phase of
each wave type. In particular, DWs are associated with
an increase in cyclonic vorticity at midlevels (~500–600
hPa) on D-1 and a subsequent increase between ~500–
3
by the optical transient detector. J. Geophys. Res.,
108, 4005, doi:10.1029/2002JD002347.
Hopsch, S. B., C. D. Thorncroft, and K. R. Tyle, 2010:
Analysis of African easterly wave structures and
their role in influencing tropical cyclogenesis. Mon.
Wea. Rev., 138, 1399–1419.
Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40year reanalysis project. Bull. Amer. Meteor. Soc.,
77, 437–471.
Kummerow, C., W. Barnes, T. Kozu, J. Shiue, and J.
Simpson, 1998: The tropical rainfall measuring
mission (TRMM) sensor package. J. Atmos.
Oceanic Technol., 15, 809–817.
Kurihara, Y., and R. E. Tuleya, 1981: A numerical
simulation study on the genesis of a tropical storm.
Mon. Wea. Rev., 109, 1629–1653.
Landsea, C. W., 1993: A climatology of intense (or
major) Atlantic hurricanes. Mon. Wea. Rev., 121,
1703–1713.
Leppert, K. D., II, and W. A. Petersen, 2010: Electricallyactive hot towers in African easterly waves prior to
tropical cyclogenesis. Mon. Wea. Rev., 138,
663−687.
Liu, Z., D. Ostrenga, G. G. Leptoukh, and A. V. Mehta,
2009: Online visualization and analysis of global
half-hourly
pixel-resolution
infrared
dataset.
th
Extended Abstracts, 25 Conf. on International
Interactive Information and Processing Systems
(IIPS) for Meteorology, Oceanography, and
Hydrology, Phoenix, AZ, Amer. Meteor. Soc., J3.4.
Molinari, J., and D. Vollaro, 2000: Planetary- and
synoptic-scale influences on Eastern Pacific tropical
cyclogenesis. Mon. Wea. Rev., 128, 3296–3307.
National Hurricane Center, cited 2011: NHC archive of
hurricane
seasons.
[Available
online
at
http://www.nhc.noaa.gov/pastall.shtml.]
Nolan, D. S., 2007: What is the trigger for tropical
cyclogenesis? Aust. Meteorol. Mag., 56, 241–266.
Raymond, D. J., S. L. Sessions, and C. Lopez Carrillo,
2011:
Thermodynamics
of
tropical
cyclogenesis in the northwest Pacific. J. Geophys.
Res., 116, D18101, doi:10.1029/2011JD015624.
Rotunno, R., and K. A. Emanuel, 1987: An air–sea
interaction theory for tropical cyclones. Part
II:
Evolutionary study using a non-hydrostatic
axisymmetric numerical model. J. Atmos. Sci., 44,
542–561.
Toracinta, E. R., D. J. Cecil, E. J. Zipser, and S. W.
Nesbitt, 2002: Radar, passive microwave, and
lightning characteristics of precipitating systems in
the tropics. Mon. Wea. Rev., 130, 802–824.
Fig. 3. As in Fig. 2, except for vertical profiles of relative
vorticity. The horizontal dashed lines indicate ± one
standard deviation.
4. CONCLUSIONS
Results suggest that the coverage by cold
cloudiness increases as D0 is approached for DWs, in
particular within the trough and southerly phases, while
the intensity of convection appears to decrease with
time in all DW phases. The evolution of NDWs also
generally shows an increase in coverage by cold
cloudiness as D0 is approached, but convective
intensity appears to remain approximately constant with
time, perhaps increasing slightly.
In addition, the
coverage by cold cloudiness is generally always greater
for DWs. In terms of large-scale wave structure, DWs
are associated with more moisture than NDWs. Hence,
results suggest that the most important characteristics
for tropical cyclogenesis and distinguishing DWs from
NDWs is a greater coverage by cold cloud tops and
enhanced large-scale moisture.
5. REFERENCES
Avila, L. A., 1991: Eastern North Pacific hurricane
season of 1990. Mon. Wea. Rev., 119, 2034−2046.
Boccippio, D. J., W. J. Koshak, and R. J. Blakeslee,
2002: Performance assessment of the optical
transient detector and lightning imaging sensor.
Part I: Predicted diurnal variability. J. Atmos.
Oceanic Technol., 19, 1318–1332.
Christian, H. J., and Coauthors, 2003: Global frequency
and distribution of lightning as observed from space
4